Generally, in coding processing of a moving picture, the amount of information is reduced by compression for which temporal redundancy and spatial redundancy in the moving picture is utilized. Generally, a method in which spatial redundancy is utilized is performed by transform into frequency domain, and a method of compression for which temporal redundancy is utilized is performed by coding using prediction between pictures (this prediction is hereinafter referred to as inter prediction). In the inter prediction coding, a current picture is coded using a coded picture preceding or following the current picture in display order as a reference picture. Next, a motion vector is derived by estimating a motion of the current picture with respect to the reference picture. Then, the difference between picture data of the current picture and prediction picture data obtained by motion compensation based on the derived motion vector is calculated to remove temporal redundancy. In the motion estimation, difference values from a current block in the current picture are calculated for blocks in the reference picture, and a block having the smallest difference value in the reference picture is used as a reference block. Then, a motion vector is derived using the current block and the reference block. There is a moving picture coding scheme called H.264 which has already been standardized (see Non Patent Literature 1).